Electronic Device ROI Tracking with Multi-Detector Fallback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Electronic devices fail to maintain the region of interest (ROI) at the center of the display screen during image capturing when the subject, such as a face, is partially covered, leading to deactivation of zooming/panning functions.
Innovation Solution
The electronic device employs multiple detectors to detect and estimate ROIs, updating the ROI position based on matching and non-matching detections across images, ensuring the ROI remains centered on the screen.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single detector is used for ROI detection, then the device complexity is reduced, but the reliability of ROI detection deteriorates when the subject is partially covered
Solution Approach 1:
The patent divides the ROI detection task into multiple specialized detectors, each responsible for detecting specific types of regions (e.g., face detector, body detector, object detector). This segmentation allows each detector to be optimized for its specific detection task, improving overall detection reliability while maintaining manageable complexity through functional specialization.
Solution Approach 2:
The patent implements a universal ROI detection system that can handle multiple detection scenarios using a set of detectors that work together. The system can detect faces, bodies, and other objects, and can switch between different detection modes based on the situation, providing multi-functional capability that improves reliability across various detection conditions.
2Reliability
If multiple detectors are used to detect ROIs, then the reliability of ROI detection is improved, but the device complexity increases
Solution Approach 1:
The patent introduces an ROI determination module that acts as an intermediary between multiple detectors and the final ROI output. This mediator receives detection results from multiple detectors, performs matching and validation, and determines the final ROI, thereby managing the complexity of multiple detectors while maintaining high detection reliability.
Solution Approach 2:
The patent implements a feedback mechanism where the ROI determination module continuously monitors detection results from multiple detectors and adjusts the detection process accordingly. When one detector fails to detect an ROI or detects an incorrect ROI, the system uses feedback from other detectors and previous detection results to correct and confirm the accurate ROI position.
3Productivity
If the ROI detection is deactivated when the subject is partially covered, then the false detection is reduced, but the image capturing functionality deteriorates
Solution Approach 1:
The patent prepares multiple detection methods and fallback mechanisms in advance to cushion against the failure of primary detection. When the primary face detector fails due to partial coverage, the system has pre-prepared alternative detection methods (body detection, object detection) ready to take over, ensuring continuous image capturing functionality without false deactivation.
Solution Approach 2:
The patent dynamically changes detection parameters and switches between different detection modes based on the detection situation. When detection accuracy drops below a threshold (indicating partial coverage), the system changes parameters such as switching from face-specific detection to more general object detection, thereby maintaining functionality while adapting to reduced precision conditions.
Data Source
AI summary
An electronic device is provided. The electronic device includes a camera module, a display, a plurality of detectors, and at least one processor, and the at least one processor may be configured to detect at least one first region of interest (ROI) in a first image received through the camera module by using a first detector for first ROI detection among the plurality of detectors, detect at least one second ROI in a second image received through the camera module by using a second detector for second ROI detection among the plurality of detectors when failing in detecting at least one first ROI matching the at least one first ROI detected in the first image in the second image by using the first detector, estimate at least one first ROI based on the at least one second ROI, update the estimated at least one first ROI to at least one first ROI when the at least one first ROI detected in the first image matches the estimated at least one first ROI, and change a position of a preview region including the updated at least one first ROI based on a position of the estimated at least one first ROI matching the at least one first ROI detected in the first image.